- Email: [email protected]
Three-dimensional carbon felt host for stable sodium metal anode
Applied Catalysis B: Environmental 259 (2019) 118113
Contents lists available at ScienceDirect
Applied Catalysis B: Environmental journal homepage: www.elsevier.com/locate/apcatb
Heterogeneous molecular rhenium catalyst for CO2 photoreduction with high activity and tailored selectivity in an aqueous solution
T
Shengbo Zhanga,b, Mei Lib, Wenjun Qiub, Jinyu Hanb, Hua Wangb, Xiao Liua,
⁎
a
Key Laboratory of Pesticide & Chemical Biology of Ministry of Education, College of Chemistry, Central China Normal University, Wuhan 430079, China Key Laboratory for Green Chemical Technology, School of Chemical Engineering and Technology, Tianjin University, Collaborative Innovation Center of Chemical Science and Engineering, Tianjin 300072, China
b
ARTICLE INFO
ABSTRACT
Keywords: CO2 photoreduction Heterogeneous molecular catalyst Organosilica nanotubes Microenvironment regulation Aqueous solution
The effective molecular-catalyst-based device that converts CO2 to value-added products is of significance but challenging. Herein, utilizing organosilica nanotubes as the solid chelating ligands, we successfully construct heterogeneous molecular rhenium catalysts for efficient CO2 photoreduction to CO. The nanotube framework compositions and the microenvironment of the rhenium catalysts are finely regulated, aiming to enhance the catalytic selectivity and activity in water-containing systems. By adjusting bipyridine amounts and capping the surface silanols, the preferential adsorption for CO2 over H2O around the active sites is realized. Relative to that of the unmodified catalyst, the CO selectivity increased greatly from 53% to 94% in water/acetonitrile. Furthermore, the heterogeneous molecular catalysts exhibit a total turnover number that is nine times higher than the corresponding homogeneous one (134 versus 15). Using in situ infrared spectroscopy and density functional theory calculations, a reasonable mechanism for high CO selectivity in the presence of H2O is demonstrated.
1. Introduction The photocatalytic reduction of CO2, which mimics natural photosynthesis, aims to achieve efficient light-to-chemical transformation [1–5]. The most serious challenge in this process is the activation of CO2 molecules because of its large potential of the CO2/CO2∙− redox couple (−1.9 V versusthe normal hydrogen electrode (NHE)). An effective alternative to activate CO2 is the proton-assisted multielectron catalytic processes, such as CO2 + 2H+ + 2e− = CO + H2O (E’ = −0.53 V), which requires much less energy [6–11]. Considering the multiple accessible redox states, transition-metal complexes could be utilized as catalysts to minimize the overpotentials for the efficient reduction of CO2. Given this, the device designs for artificial photosynthesis (AP) based on molecular catalysts have always been attracted much research attention [12–15]. However, the molecule-assembled AP systems usually perform low stability and activity. Furthermore, photocatalytic CO2 reduction using molecular catalysts is usually confined to non-aqueous solvent (CH3CN or DMF) systems, which is mainly because that reducing a proton to H2 is more thermodynamically favorable than reducing CO2 in water [16–19]. Given that a water-containing system is necessary to construct AP devices that couples CO2 reduction and water oxidation, designing and developing molecular catalysts with
⁎
high selectivity and stability for the photocatalytic CO2 reduction in a water-containing system are of high significance. The incorporation of metal complex catalysts onto a solid matrix, such as metal organic frameworks (MOFs), covalent organic frameworks (COFs) and periodic mesoporous organosilicas (PMOs) might be a promising way to design AP devices for efficient CO2 photoreduction [20–28]. On the one hand, heterogenization can make active sites spatial isolation and prevent the molecular catalyst deactivation. On the other hand, the matrix framework compositions could be adjusted to preferentially adsorb CO2 rather than water molecules, thus tailoring the catalytic activity and selectivity in water-containing systems. Recently, chelating-ligand-incorporated organosilica nanotubes were prepared from bridged organosilane precursors for the efficient heterogeneous molecular catalyst design in our group [29–34]. These nanotubes with mesoporous diameters have distinct advantages, including various organic functionalities in the nanotube frameworks, high surface areas, fast diffusion of substrates in the tubes, and tunable hydrophilicity/hydrophobicity. Herein, using benzene-bridged organosilica nanotubes containing bipyridine ligands (BPy-NT) as the platform, we report the direct construction of a rhenium complex on the nanotube walls (Re–BPy-NT) through metal complexation for photocatalytic reduction of CO2. By carefully adjusting the bipyridine
Corresponding author. E-mail address: [email protected] (X. Liu).
https://doi.org/10.1016/j.apcatb.2019.118113 Received 14 June 2019; Received in revised form 14 August 2019; Accepted 22 August 2019 Available online 23 August 2019 0926-3373/ © 2019 Elsevier B.V. All rights reserved.
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
4800 scanning electron microscope (SEM, 5 kV) equipped with the Thermo Scientific energy-dispersion X-ray fluorescence analyser. N2 adsorption-desorption isotherms were tested at −196 °C on a Micromeritics Tristar 3000 instrument. The specific surface area was obtained by the Brunauer–Emmett–Teller method; the pore size distribution was obtained by the Barrett–Joyner–Halenda method using the adsorption branch. Carbon, hydrogen and nitrogen contents were determined viaCHN elemental analysis (vario EL CUBE). UV/vis absorption and diffuse reflectance spectra were obtained using Instant Spec BWS003 spectrometers. Re LIII-edge extended X-ray absorption fine structure (EXAFS) spectra were collected at the beamline 1W1B of Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences. The beamline was equipped with a Si (111) double crystal monochromator. The k3-weighted EXAFS oscillations were Fourier transformed into R-space. Transient absorption spectroscopy (TAS) was carried out employing a Femtosecondscale laser flash photolysis spectrometer (Applied Photophysics) equipped with the proper diffuse reflectance accessory. The samples were placed in a closed quartz cuvette, purged with N2 and were excited with a 380 nm laser beam. FT-IR spectra were collected with a Bruker Vertex 80v. In situ IR spectra of a acetonitrile-water-triethanolamine (0.875:1:0.125, v/v) solution containing Re catalyst during photoirradiation (λ > 420 nm) under a CO2 atmosphere. Solid-state 13C cross polarization magic-angle spinning (CP MAS) NMR and 29Si magic angle spinning (MAS) NMR spectra were collected on 400 MHz instruments. Photoluminescence spectra (PL) were done on Jobin Yvon Fluorolog3-21 spectrophotometer at an excitation wavelength of 380 nm by using the powder sample. Photoelectric properties measurements were carried out with an Autolab 302 N electrochemical system with a 300 W xenon lamp and used an Ag/AgCl (0.5 M Na2SO4) as the reference electrode and a Pt plate as the counter electrode. The photocurrent density vs time (I-t) curves of the prepared photoelectrodes were operated under an alternate light irradiation (light on/off cycles: 30 s). Mott–Schottky measurements of the catalysts were conducted in an Autolab 302 N electrochemical system at frequencies of 500, 1000, and 1500 Hz. In Mott–Schottky measures, a three-electrode configuration was used with Pt plate as counter electrode, Ag/AgCl electrode as reference electrode and catalysts on FTO glass substrate as working electrode in 0.5 M Na2SO4 electrolyte. Water contact angles were measured on the contact angle system OCA 20 by pressing the powder sample into a thin sheet.
Scheme 1. Synthetic routes for the design of heterogeneous molecular Rebased catalysts on organosilica nanotubes (Re-BPy0.3-NT-Me).
contents in the framework, the heterogeneous molecular-catalyst-based nanotubes (Re–BPy0.3–NT) could not only facilitate the rapid diffusion of reactants and products but also increase the adsorption capacity of CO2. In order to further inhibit protons from reaching the active sites to produce H2, the remaining accessible surface silanols groups (Si−OH) of the organosilica nanotubes were capped using trimethylsilyl [–Si (CH3)3] (Re–BPy0.3–NT–Me) to increase the hydrophobicity (Scheme 1). The as-prepared Re–BPy0.3–NT–Me shows higher catalytic activity, CO selectivity, and a turnover number (TON) compared with those of the unmodified catalyst and even the corresponding homogeneous one in the CO2 photoreduction system with a high concentration of water. 2. Experimental section 2.1. Synthesis of Re-BPy0.3-NT-Me The rhenium-immobilized nanotube catalysts were prepared by adding BPy-NT (300 mg) to a solution of Re(CO)5Cl (40 mg) in 60 mL of toluene under nitrogen atmosphere. After the suspension was stirred under refluxing conditions for 24 h, the solid phase was obtained by filtration and washed with acetone. The dried samples were named as Re-BPy0.3-NT. Then, 2.0 g of Re-BPy0.3-NT and 8 mL of [(CH3)3Si)]2NH in 30 mL of dry toluene were stirred at room temperature overnight. The resultant powder was filtered, washed and dried under vacuum overnight at 60 °C. The dried samples were named as Re-BPy0.3-NT-Me.
2.4. DFT theoretical calculation All the calculations were used spin-polarized density functional theory (DFT) with generalized gradient approximation (GGA) for exchange-correlation potential embedded in the Vienna Ab Initio Simulation Package (VASP). The ion-electron interaction was described with projector augmented wave (PAW) method. The cutoff energy of the plane-wave basis set was set at 500 eV. The cluster with periodic boundary condition was modeled with a large super cell of 28 × 28 × 28 Å3, eliminating the artificial interaction between molecular electrocatalysts in adjacent cells. The van der Waals (vdW) dispersion by employing the D3 method of Grimme was considered for all the calculation. The Brillouin zone integrations were performed by using Gamma 1 × 1×1 for geometric optimization. The convergence thresholds for structural optimization was set at 0.02 eV/Å in force. The convergence criterion for energy was 10−5 eV.
2.2. Re-BPy0.3-NT-Me-photocatalyzed CO2 reduction reaction In a typical experiment, Re-BPy-NT-Me (2.9 μmol Re) was added to a fresh 40 mL mixed solution containing 17.5 mL of acetonitrile, 20 mL of water and 2.5 mL of triethanolamine. This system was purged with CO2 for 1.0 h. The reaction was operated at the room temperature with a visible light (λ > 420 nm) illumination. The gas molecules generated in the head space were periodically analyzed quantitatively by a Bruker 456 gas chromatograph equipped with a thermal conductivity detector (TCD).
3. Results and discussion
2.3. Characterizations
A rhenium complex was prepared, as shown in Scheme 1. First, organosilica nanotubes containing bipyridine ligands in the framework (BPyx-NT, where x corresponds to the molar ratio of precursors 2 in the initial total precursors) was synthesized as our previous report [29,33]. Subsequently, the rhenium complex [Re(CO)5Cl] was anchored on the
Transmission electron microscopys (TEM) were operated on a JEM2100 F system operating at 200 kV. The morphology of these catalysts were analysed by scanning electron microscopy (SEM) using Hitachi S2
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 1. (a) TEM, (b) SEM and (c) EDS mapping images of Re–BPy0.3–NT–Me. The inset graph in (a) shows the nanotube length distribution in ethanol obtained viadynamics light scattering and magnified high-resolution TEM. (d) Aberration-corrected HAADF-STEM and (e–f) magnified aberration-corrected HAADF-STEM and corresponding Re mapping images of Re–BPy0.3–NT–Me.
nanotube walls (denoted by Re–BPy0.3–NT) viathe metal complexation of Re to the bipyridine unit in BPy0.3–NT (Figs. S1–5). Finally, the remaining accessible silanols (Si–OH) on the organosilica nanotubes were capped using trimethylsilyl [–Si(CH3)3] groups. The obtained sample was denoted by Re–BPy0.3–NT–Me. For comparison, the Re complex was also incorporated onto bipyridine-bridging organosilica nanotubes with fewer bipyridine groups, i.e., BPy0.1–NT (denoted by Re–BPy0.1–NT) (Figs. S7–11) and pure silica nanotubes without phenyl groups (denoted by Re–BPy0.1–SNT) (Figs. S13–17). Fig. 1a and b show the transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images of the short nanotubes Re–BPy0.3–NT–Me. The TEM image clearly reveals that these catalysts comprise nanotubes with ˜60 nm lengths and ˜5-7 nm inner diameters. The SEM image further confirms the nanorod morphology. The energydispersive X-ray spectrometry (EDS) mapping analysis reveals that Re is uniformly distributed on the support (Fig. 1c). In order to more intuitively know the distribution of Re atoms, we performed aberrationcorrected, high-angle, annular, dark-field scanning TEM (HAADFSTEM) measurements (Fig. 1d–f) with sub-angstrom resolutions. Two regions clearly appear in these images; broad, curved, dark lines and small bright spots sized less than ˜0.5 nm. The dark regions correspond to the projections of the nanotube walls along the incident electron beam direction, whereas the bright spots indicate that isolated heavier atoms were present, which were Re atoms in this case. Since the bright spots are not concentrated on the outer/inner surfaces of the nanotube walls but are rather uniform throughout the walls, one can conclude that the Re atoms were uniformly distributed along the nanotube walls with single-site states. The inductively coupled plasma (ICP) measurements show that the Re loadings were 0.288, 0.286, 0.294, and 0.291 mmol g−1 for Re–BPy0.1–SNT, Re–BPy0.1–NT, Re–BPy0.3–NT, and Re–BPy0.3–NT–Me, respectively (Table S1). According to the CHN elemental analysis, the bpyWithout coordination/bpyTotal molar ratios for each catalyst were determined to be 42%, 38%, 73%, and 75%, respectively. To further demonstrate the geometric structure of molecular Re on the skeleton of organosilica nanotubes, various spectroscopies were performed on Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl. In contrast with that of BPy0.3–NT, the Fourier transform infrared (FT-IR) spectrum (Fig. 2a) of Re–BPy0.3–NT–Me exhibits two new absorption peaks at approximately 1921 and 2028 cm−1, which were assigned to the vibration of the Re–CO ligands, in accordance with that of homogeneous Re(bpy)(CO)3Cl. The UV/vis diffuse reflectance spectrum (Fig. 2b) of
Re–BPy0.3–NT–Me shows a new absorption band appeared at approximately 400 nm compared with that of BPy0.3–NT, similar to that of homogeneous Re(bpy)(CO)3Cl, which could be assigned to the metal-toligand (i.e., Re-to-bipyridine-ligand) charge transfer (MLCT) [14,24,26]. The X-ray absorption near-edge structure spectroscopies (XANES) show that the white line resonance peak at approximately 10,538 eV, which is assigned to the allowed 2p–5d transition, intensified with the increase in the d-band vacancies resulting from oxidation (Fig. 2c). Thus, the white line absorption peaks of oxidized Re species are more intense than those of the reduced species, which agrees well with the XPS results. Furthermore, more detailed structural information was obtained viaextended X-ray absorption fine structure spectroscopy (EXAFS). In the Fourier transform of the Re LIII-edge extended EXAFS spectra of Re(bpy)(CO)3Cl and Re–BPy0.3–NT–Me (Fig. 2d), three strong peaks appear at approximately 0.165, 0.220, and 0.264 nm, which can be assigned to Re–C, Re–N, and Re–Cl bond lengths, respectively [14,26]. The XANES spectra and EXAFS Fourier transform at the Re LIIIedge show spectral features corresponding to those of the complex Re (bpy)(CO)3Cl, suggesting the successful formation of the desired Re complex on the pore surface of BPy-NT. To prove the hydrophobicity/hydrophilicity of the as-prepared catalysts, water contact angles were measured (Fig. 3a). Organosilica nanotubes with benzene groups in the skeleton exhibit a significantly higher contact angle and thus higher hydrophobicity than pure silicon nanotubes. More importantly, Re–BPy0.3–NT–Me end-capped using trimethylsilyl [–Si(CH3)3] groups shows the strongest hydrophobicity, with a water contact angle of 96°. These results indicate that hydrophilicity/hydrophobicity can be regulated by finely tuning the matrix framework compositions. In addition, gas sorption using water and benzene was measured at 298 K. As shown in Fig. 3b, c, Re–BPy0.1–NT and Re–BPy0.3–NT exhibit lower water adsorption and higher benzene adsorption at P/P0 = 0.97 than Re–BPy0.1–SNT. Clearly, Re–BPy0.3–NT–Me shows the lowest water adsorption (35.3 cm3/g) and highest benzene adsorption (138.9 cm3/g) at P/P0 = 0.97 because of the improved hydrophobicity. In addition, the CO2 uptake performances of all the materials were investigated at 298 K (Fig. 3d). The adsorption capacities of Re–BPy0.3–NT–Me and Re–BPy0.3–NT were 13.2 and 10.7 mL/g, respectively, which were significantly higher than those of Re–BPy0.1–NT and Re–BPy0.1–SNT. The higher CO2 adsorption capacities of 3
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 2. (a) FT-IR and (b) UV/vis diffuse reflectance spectra of BPy0.3–NT, Re–BPy0.3–NT–Me, and Re(bpy)(CO)3Cl. (c) Re Lш-edge XANES spectra, and (d) Fourier transform EXAFS spectra of Re(bpy)(CO)3Cl, Re–BPy0.3–NT–Me and Re–BPy0.3–NT–Me after reaction.
Fig. 3. (a) Contact angle measurements of Re–BPy0.1–SNT, Re–BPy0.1–NT, Re–BPy0.3–NT and Re–BPy0.3–NT–Me at 298 K. (b) Water and (c) benzene sorption measurements and (d) CO2 uptake performance of Re–BPy0.1–SNT, Re–BPy0.1–NT, Re–BPy0.3–NT, and Re–BPy0.3–NT–Me at 298 K. (e) In situ infrared of CO2 for pure 2,2'-bipyridine at 298 K.
4
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 4. (a) PL spectra, (b) Transient photocurrent responses, (c), (d) Transient optical spectra in the presence of a sacrificial electron donor TEOA, (e) Comparison of kinetics for intramolecular charge transfer at 550 nm and (f) Mott-Schottky plots of Re(bpy)(CO)3Cl and Re–BPy0.3–NT–Me.
Re–BPy0.3–NT–Me and Re–BPy0.3–NT could be primarily attributed to the more bipyridine ligands in their frameworks, which can effectively improve the CO2 uptake (Fig. 3e, Table S1). The optical properties of Re–BPy0.3–NT–Me were investigated by photoluminescence (PL) spectroscopy, as shown in Fig. 4a. Re–BPy0.3–NT–Me exhibits a lower PL intensity than homogeneous Re(bpy) (CO)3Cl when irradiated by a 380-nm laser, with a peak centered at 575 nm, which can be attributed to the efficient photoinduced charge transfer process resulting from the inhibition of intermolecular aggregation by the heterogenization of the catalyst. The transient photocurrent response was also performed to study the photogenerated charge transfer and separation efficiency (Fig. 4b). The photocurrent immediately increased when the light irradiation was turned on, while the photocurrent quickly decreased when the light was switched off, which demonstrates the existence of the photoresponse for Re–BPy0.3–NT–Me. The heterogeneous molecular Re–BPy0.3–NT–Me shows a photocurrent density of approximately 5.5 μA/cm2, which is higher than the complex Re(bpy)(CO)3Cl (3.0 μA/cm2), indicating the efficient photogenerated charge transfer and separation in Re–BPy0.3–NT–Me. To further study light absorption and intramolecular charge transfer (ICT) of Re–BPy0.3–NT–Me, the transient absorption (TA) spectroscopies were performed with a 380 nm excitation (Fig. 4c, d). A stronger and broader positive transient signal in the 550–700 nm region assigned to the absorption of excited ICT appears for Re–BPy0.3–NT–Me compared with the complex Re(bpy)(CO)3Cl [20], which may be mainly due to the fact that the heterogeneous molecular complex has longer lifetime in the excited ICT states than Re(bpy)(CO)3Cl. The TA of Re(bpy)(CO)3Cl shows a positive transient signal in the > 725 nm region, which can be assigned to the formation of catalytically inactive Re-dimers [28], whereas Re–BPy0.3–NT–Me did not have this transient signal due to isolated active sites. Furthermore, the corresponding kinetics comparison of ICT for Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl were also performed. As shown in Fig. 4e, the formation (rising component) of the excited ICT is ultrafast with ˜300 fs for both samples. But interestingly, this excited ICT state of Re–BPy0.3–NT–Me exhibits much longer lifetime (t =161.5 ps) than that of the complex Re(bpy)(CO)3Cl (t = 24.8 ps), suggesting that the heterogenization of Re(bpy)(CO)3Cl
could effectively inhibit charge recombination in Re–BPy0.3–NT–Me. In addition, Mott-Schottky measurements were applied to investigate the electronic band positions and corresponding redox abilities of Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl (Fig. 4f). The flat band positions are −1.68 and −1.50 V (vs Ag/AgCl) for Re–BPy0.3–NT–Me and Re(bpy) (CO)3Cl, respectively, according to extrapolation of C2− values at frequencies of 500, 1000, and 1500 Hz. The positive slopes reveal n-typelike conductivities of Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl. For ntype semiconductors, the flat band potential is generally approximated by the conduction band potential. Obviously, both samples show more negative reduction potentials (−1.48 V vs NHE for Re–BPy0.3–NT–Me and −1.30 V vs NHE for Re(bpy)(CO)3Cl) than CO2/CO (−0.53 V vs NHE), suggesting the feasibility for photoreduction of CO2 to CO. Subsequently, the performances (Fig. S18) of these Re-based catalysts for the photochemical reduction of CO2 were investigated at room temperature with visible light (λ > 420 nm) illumination in a CO2–saturated acetonitrile–water solution (VMeCN/VH2O = 0.875/1). Fig. 5a shows the CO-TON, H2-TON and CO selectivity with different catalysts under similar reaction conditions at 8 h. No other liquid phase products were detected (Fig. S19). The CO-TON (15.20) of Re–BPy0.3–NT was higher than that of Re–BPy0.1–NT (10.72). The higher catalytic activity of Re–BPy0.3–NT could be attributed to a higher number of bipyridine ligands in the framework (increase CO2 adsorption) and the shorter nanotube length (decrease the diffusion path of reactants and products during the reactions). Notably, the Re–BPy0.3–NT–Me with the capped silanol exhibited the highest CO-TON (19.30) and lowest H2-TON (1.23) with the CO producing rate of 700 μmol/g/h. A CO selectivity as high as 94% could be achieved even in the high water-containing system (50% volume ratio, Table 1). This high selectivity can be explained by the precise modulation of the surface physiochemical properties. When the silanol was capped by methyl groups, the high hydrophobicity of Re–BPy0.3–NT–Me could effectively prevent H2O molecules from reaching the catalytic active centers for H+ reduction. In contrast, CO2 molecules can easily approach the Re sites, thus facilitating the CO2 reduction owing to the enhanced role of CO2 in the competition for adsorption sites. To further prove the hydrophobic/hydrophilic effect, the catalyst with pure silica 5
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 5. (a) The CO-TON, H2-TON and CO selectivity of Re–BPy0.1-SNT, Re(bpy)(CO)3Cl, Re–BPy0.1–NT, Re–BPy0.3–NT and Re– BPy0.3–NT–Me after 8 h at 298 K with 20 mL of water and 2.9 μmol Re. (b) Dependence of the QYs of the Re–BPy0.3–NT–Me on the wavelength of monochromatic light irradiation (435, 500 nm) QY = 2nco/nincident photons × 100%. (c) Recycling performance, (d) The total CO-TON and (e) filtration experiment of Re–BPy0.3–NT–Me for the photochemical reduction of CO2 with 1.0 μmol Re. (f) Effect of water on the CO-TON, H2-TON and CO selectivity for the Re–BPy0.3–NT–Me and Re–BPy0.1–SNT at 298 K with 2.9 μmol Re for the photochemical reduction CO2 (the total volume of solution is 40 mL).
Re(bpy)(CO)3Cl was also used to photocatalyze the same reaction. Unfortunately, Re(bpy)(CO)3Cl became inactive after 4 h of reaction (Fig. S18) and showed a lower CO selectivity of 70%, owing to the decomposition via bimolecular pathways to form metal nanoparticles (Figs. S20, S28) [22,25,27,29,32]. These above results demonstrated the excellent catalytic performance of Re–BPy0.3–NT–Me for the CO2–CO reaction owing to the heterogenization of the molecular catalyst and the finely tuned microenvironment around the active species. In addition, the quantum yields (QYs) strongly depend on the intensity of irradiated light (Fig. 5b). In particular, the QYs of Re–BPy0.3–NT–Me can be as high as 0.146% under 435-nm monochromatic light irradiation. After the reaction, Re–BPy0.3–NT–Me was reused for the next run, and the solid catalyst retained high catalytic activity (Fig. 5c). During the fifth recycle, CO-TON was maintained at approximately 22.85, showing a slight decrease. The total TON reached 134 within 40 h (Fig. 5d). In contrast, the homogeneous catalyst showed a large decrease in CO-TON for the first recycle and was fully deactivated after the second reuse (15 of the total TON, Fig. 5d). In addition, the filtration experiment indicated that the catalytic activity came from Re–BPy0.3–NT–Me, and the solid catalyst was truly heterogeneous (Fig. 5e). The structure analysis of Re-BPy0.3-NT-Me after the reaction revealed the molecular nature of the active site without the formation of Re nanoparticles, as well as the stable nanotube frameworks (Fig. 2c, d, Figs. S20–27). To demonstrate the effect of hydrophobicity/hydrophilicity on the reaction pathways, the effect of the amount of water on the CO-TON,
Table 1 Comparison of catalyst properties during the photochemical reduction CO2.a Entry
Photocatalyst
H2-TONb
CO-TONc
Selectivity (%)
1 2 3 4 5 6 7 8 9 10 11 12
Re-BPy0.3-NT-Me Re-BPy0.3-NT Re-BPy0.1-NT Re-BPy0.1-SNT Re(bpy)(CO)3Cl – BPy0.3-NT Re-BPy0.3-NT-Me/ no CO2 Re-BPy0.3-NT-Me/ no light Re-BPy0.3-NT-Me/no TEOA Re-BPy0.3-NT-Me/2.0 μmol Re Re-BPy0.3-NT-Me/1.0 μmol Re
1.23 2.44 3.16 5.56 4.18 0 0 1.41 0 0 1.48 1.94
19.30 15.20 10.72 6.32 9.64 0 0 0 0 0 24.21 30.60
94 86 77 53 70 – – 0 – – 94 94
a
Reaction conditions: 2.9 μmol of Re, 2.5 mL of TEOA, 17.5 mL of CH3CN, 20 mL of H2O, and λ > 420 nm, at room temperature. b H2-TON is defined as the number of evolved hydrogen molecules per catalytic site (Re). c CO-TON is defined as the number of evolved CO molecules per catalytic site (Re).
nanotubes but no phenyl groups (Re–BPy0.1–SNT) was used to photocatalyze the CO2–CO reaction under the same reaction conditions. It can be seen that Re–BPy0.1–SNT exhibited the lowest CO-TON (6.32) and highest H2-TON (5.56), with a CO selectivity of 53% (Fig. 5a), which could be mainly attributed to the high hydrophilicity of the pure silica nanotube framework. For comparison, the homogeneous catalyst 6
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 6. (a) CVs of Re–BPy0.3–NT–Me in a 0.1-M NBu4PF6 CH3CN/H2O (v/v = 0.875:1) solution with 1 mM Re catalysts under a saturated N2 (black) and CO2 (red) atmosphere using a 3mm glassy carbon electrode and Ag/AgCl electrode at a scan rate of 100 mV s−1. (b) In situ FT-IR spectra of CO2 photoreduction over Re–BPy0.3–NT–Me recorded after various irradiation times.
the ligand exchange of the molecular CO of species E with Cl− produces the initial Re complex A to complete the CO2 reduction pathway (1). Evidently, adding water plays a proton-promoting role. Simultaneously, protons are reduced to form H2 (pathway (2)) during this process, which is the main reaction competing with the CO2 reduction [17–19,45]. Interestingly, for hydrophobic Re–BPy0.3–NT–Me, more CO2 and fewer protons are allowed to reach the active site to react; thus, water primarily play a proton-promoting role. However, for hydrophilic Re–BPy0.1–SNT, excess protons can reach the active sites, resulting in more competition between the two reactions (1 and 2), thus weakening the reduction of CO2 (1). In addition, density functional theory (DFT) calculations were employed to investigate the binding energies of H2O onto the Re–BPy0.3–NT and Re–BPy0.3–NT–Me. As shown in Fig. 7b, the binding energy of H2O on Re–BPy0.3–NT is −0.48 eV, which is much stronger than that of Re–BPy0.3–NT–Me (−0.28 eV). Further, such an obvious difference in the binding energy of H2O over Re–BPy0.3–NT and Re–BPy0.3–NT–Me has been proved by experimentally (Fig. 3a, b, and c). Notably, the weaker binding energy of H2O on Re–BPy0.3–NT–Me can suppress the undesired metal–H2O adduct F with H2 as a by-product. Finally, the evolution of the free energy of CO2 photoreduction to CO on the Re–BPy0.3–NT–Me was calculated using DFT to better evaluate and understand the reaction processes and related active species (Fig. 7c). As shown in Fig. 7c, the pathway for the reduction of CO2 to CO viaa Re–carboxyl (Re–COOH*) intermediate viathe transfer of two protons and electrons was considered. Initially, the CO2 molecule is adsorbed on the Re sites, with concerted protonation and electron transfer, leading to the formation of the Re–COOH* intermediate. Subsequently, this intermediate can be further reduced viaproton–electron transfer to form a Re–CO* adduct, releasing one H2O molecule. Here, it should be mentioned that the hydrophobic material Re–BPy0.3–NT–Me facilitates the separation of water from the reaction system, thereby accelerating the reaction. The main product CO is finally generated from the ligand exchange of the molecular CO of species E with Cl−. We find that the formation of the Re–COOH* species viaprotonation (energy barrier of 0.39 eV) and the release of molecular CO viathe ligand exchange (energy barrier of 0.46 eV) are the two main rate-limiting steps. Importantly, adding weak Bronsted acids (H2O) can effectively facilitate these catalytic pathways. On the one hand, adding water provides more protons to accelerate the formation of the Re–COOH* intermediate viathe protonation of CO2. On the other hand, the protons H+ from the H2O can further promote the conversion of ReOH* to Re–CO*. As a result, the concentration of Re–CO* will continue to increase, thereby accelerating the release of CO according to reaction equilibrium theory. However, the DFT calculations of the free energy evolution of protons reducing to H2 indicate that reducing protons is easier than reducing CO2 in this system (Fig. S31), which is the main competitive reaction for CO2 reduction. In response to this problem, adjusting the microenvironment around the active sites will provide
H2-TON and CO selectivity in this reaction system was investigated using Re–BPy0.3–NT–Me and Re–BPy0.1–SNT as control catalysts (Fig. 5f). Surprisingly, adding water can improve the catalytic performance of the hydrophobic Re–BPy0.3–NT–Me catalyst to some extent, mainly due to that the added water could promote the proton-coupled electron transfer [35–40]. Furthermore, the presence of the protoncoupled electron transfer effect was also confirmed by the photocurrent response test (Fig. S29). A high CO selectivity over 94% can be obtained with a high CO-TON over 18 and an almost constant H2-TON below 1.5 in 0–20 mL of water. However, for the hydrophilic Re–BPy0.1–SNT, the CO selectivity and CO-TON always decrease, whereas H2-TON increases with the addition of water. These results indicate that the hydrophobic Re–BPy0.3–NT–Me catalyst can achieve high activity and selectivity in a system with a high concentration of water by fine-tuning the microenvironment around the active species, which provides a promising platform to design devices for artificial photosynthesis. Next, the redox properties of Re–BPy0.3–NT and Re–BPy0.3–NT–Me were investigated by recording cyclic voltammograms (CVs) to further illustrate the effect of hydrophobicity on CO2 reduction (Fig. 6a and Fig. S30). The CV of Re–BPy0.3–NT–Me exhibits increased current at Eonset = −0.74 V under a CO2 atmosphere, and this current density (−1.22 mA/cm2) is more than 2.3 times higher than that under an N2 atmosphere. Interestingly, Re–BPy0.3–NT–Me exhibits a lower Eonset and higher current density than Re–BPy0.3–NT (−0.80 V and −0.94 mA/ cm2, respectively) and Re(bpy)(CO)3Cl (−0.82 V and −0.82 mA/cm2, respectively). These results indicate that hydrophobic Re–BPy0.3–NT–Me is more favorable for CO2 reduction. To further investigate the photocatalytic CO2 reduction process, in situ infrared spectroscopy was conducted for the Re–BPy0.3–NT–Me sample (Fig. 6b). Some new peaks clearly emerge with the increased irradiation time from 0 to 64 min; the peaks at 1297, 1473, and 1650 cm−1 are assigned to the Re–COOH groups [41–44]. The intensity of these peaks increased with increasing exposure time, suggesting that the Re–COOH groups are intermediates during the photoreduction of CO2 to CO. Subsequently, the Re–COOH species accepts another proton and undergoes dehydration to form a Re–CO complex, which gives rise to some additional peaks at 1870–2050 cm−1. Finally, the CO molecule in Re–CO can be readily released and desorbed from the catalyst surface, as indicated by the peak at 2160 cm−1 [41–44]. Based on the experimental results, a possible mechanism underlying the photocatalytic reduction of CO2 in a system containing such a high concentration of water is proposed in Fig. 7a. First, the initial Re complex A is converted into the one-electron reduced species B after irradiation with visible light based on the MLCT. Then, species B is reductively quenched using TEOA to form species C. Subsequently, the CO2 adduct D is formed by the nucleophilic attack of the Re complex C on the electrophilic carbon atom of CO2, accompanied by coordination with H+ of H2O. The CO2 adduct D can accept another H+ from H2O to form intermediate species E, thus driving the CO2 reduction. Finally, 7
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 7. (a) Possible photocatalytic CO2 reduction mechanism. DFT calculations: (b) the binding energy of H2O onto Re–BPy0.3–NT and Re–BPy0.3–NT–Me and (c) free energy evolution of the reduction of CO2 to CO on Re–BPy0.3–NT–Me.
obvious advantages. Furthermore, it must be mentioned that the desorption of CO intermediates on the molecular catalyst is also very important for photoreduction CO2 to CO [13,18,20]. Through the hydrophobicity regulation, we hope to reduce the competitive reaction of water, and also hope not to reduce the desorption process of the product CO. Fortunately, the DFT results show that binding energy of ReCO on the Re-BPy0.3-NT-Me and Re-BPy0.3-NT is −1.15 eV and −1.08 eV (Fig. S32), respectively, which indicates that the desorption of product CO on the molecular catalyst is relatively similar. Therefore, we believe that the improvement of CO2 reduction performance for the modified Re-BPy0.3-NT-Me material is mainly by reducing the competitive reaction of H2O according to the DFT and experimental results. For the optimized Re–BPy0.3–NT–Me, such a microenvironment can not only prevent excessive protons from reacting on the active site owing to high hydrophobicity but also effectively increase the adsorption capacity of CO2 owing to more bipyridine on the skeleton, thus playing a dual role.
intermediate was Re–COOH* viathe protonation of CO2 after modifying the microenvironment around the active sites. We believe that this work will provide a new guideline for the device design using functional organosilica nanotubes as highly tunable one-dimensional catalytic materials for artificial photosynthesis. Declaration of Competing Interest The authors declare no competing financial interests. Acknowledgments We acknowledge the National Natural Science Foundation of China (U1662109) for financial support. The Program of Introducing Talents of Discipline to Universities of China (111 program, < GN2 > B17019) < /GN2 > is also acknowledged. Appendix B. Supplementary data
4. Conclusion
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.apcatb.2019.118113.
In conclusion, we have constructed efficient heterogeneous molecular rhenium catalysts based on organosilica nanotubes and proposed a microenvironment modification strategy to tune the surface compositions for the highly selective photoreduction of CO2 in aqueous solutions. By adjusting the bipyridine contents in the framework of organosilica nanotubes and subsequently capping the surface silanols, we modulated the CO2 adsorption capacity and hydrophobicity/hydrophilicity around the active sites, leading to a high CO selectivity of 94% in a photocatalytic CO2 reduction system with a high concentration of water. DFT calculations and in situ infrared spectroscopy reveal that the binding energy of H2O decreased considerably, and the main
References [1] X. Liu, S. Inagaki, J. Gong, Angew. Chem. Int. Ed. 55 (2016) 14924–14950. [2] N. Elgrishi, M.B. Chambers, X. Wang, M. Fontecave, Chem. Soc. Rev. 46 (2017) 761–796. [3] S. Berardi, S. Drouet, L. Francas, C. Gimbert-Surinach, M. Guttentag, C. Richmond, T. Stoll, A. Llobet, Chem. Soc. Rev. 43 (2014) 7501–7519. [4] X. Chang, T. Wang, J. Gong, Energy Environ. Sci. 9 (2016) 2177–2196. [5] X. Liu, F. Kang, C. Hu, L. Wang, Z. Xu, D. Zheng, W. Gong, Y. Lu, Y. Ma, J. Wang, Nat. Chem. 10 (2018) 1201–1206. [6] M. Aresta, A. Dibenedetto, A. Angelini, Chem. Rev. 114 (2014) 1709–1742.
8
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al. [7] E.E. Benson, C.P. Kubiak, A.J. Sathrum, J.M. Smieja, Chem. Soc. Rev. 38 (2009) 89–99. [8] K.A. Grice, Coord. Chem. Rev. 336 (2017) 78–95. [9] A.J. Morris, G.J. Meyer, E. Fujita, Acc. Chem. Res. 42 (2009) 1983–1994. [10] M.L. Clark, A. Ge, P.E. Videla, B. Rudshteyn, C.J. Miller, J. Song, V.S. Batista, T. Lian, C.P. Kubiak, J. Am. Chem. Soc. 140 (2018) 17643–17655. [11] B. Reuillard, K.H. Ly, T.E. Rosser, M.F. Kuehnel, I. Zebger, E. Reisner, J. Am. Chem. Soc. 139 (2017) 14425–14435. [12] K.M. Choi, D. Kim, B. Rungtaweevoranit, C.A. Trickett, J.T. Barmanbek, A.S. Alshammari, P. Yang, O.M. Yaghi, J. Am. Chem. Soc. 139 (2017) 356–362. [13] B. Zhang, L. Sun, Chem. Soc. Rev. 48 (2019) 2216–2264. [14] M. Waki, K.I. Yamanaka, S. Shirai, Y. Maegawa, Y. Goto, Y. Yamada, S. Inagaki, Chem. Eur. J. 24 (2018) 3846–3853. [15] R. Kuriki, M. Yamamoto, K. Higuchi, Y. Yamamoto, M. Akatsuka, D. Lu, S. Yagi, T. Yoshida, O. Ishitani, K. Maeda, Angew. Chem. Int. Ed. 56 (2017) 4867–4871. [16] S. Fukuzumi, Y.M. Lee, H.S. Ahn, W. Nam, Chem. Sci. 9 (2018) 6017–6034. [17] S. Sato, T. Morikawa, T. Kajino, O. Ishitani, Angew. Chem. Int. Ed. 52 (2013) 988–992. [18] T. Ouyang, H.-H. Huang, J.-W. Wang, D.-C. Zhong, T.-B. Lu, Angew. Chem. Int. Ed. 56 (2017) 738–743. [19] M. Liu, Y.-F. Mu, S. Yao, S. Guo, X.-W. Guo, Z.-M. Zhang, T.-B. Lu, Appl. Catal. B 245 (2019) 496–501. [20] S. Yang, W. Hu, X. Zhang, P. He, B. Pattengale, C. Liu, M. Cendejas, I. Hermans, X. Zhang, J. Zhang, J. Huang, J. Am. Chem. Soc. 140 (2018) 14614–14618. [21] Y. Wang, N.Y. Huang, J.Q. Shen, P.Q. Liao, X.M. Chen, J.P. Zhang, J. Am. Chem. Soc. 140 (2018) 38–41. [22] X. Liu, Y. Maegawa, Y. Goto, K. Hara, S. Inagaki, Angew. Chem. Int. Ed 55 (2016) 7943–7947. [23] G. Jiang, K. Geng, Y. Wu, Y. Han, X. Shen, Appl. Catal. B 227 (2018) 366–375. [24] H.-P. Liang, A. Acharjya, D.A. Anito, S. Vogl, T.-X. Wang, A. Thomas, B.-H. Han, ACS Catal. 9 (2019) 3959–3968. [25] Y. Ueda, H. Takeda, T. Yui, K. Koike, Y. Goto, S. Inagaki, O. Ishitani, ChemSusChem 8 (2015) 439–442. [26] M. Wen, K. Mori, Y. Kuwahara, H. Yamashita, ChemCatChem 7 (2015) 3519–3525.
[27] C. Wang, Z. Xie, K.E. deKrafft, W. Lin, J. Am. Chem. Soc. 133 (2011) 13445–13454. [28] C.D. Windle, E. Pastor, A. Reynal, A.C. Whitwood, Y. Vaynzof, J.R. Durrant, R.N. Perutz, E. Reisner, Chem. Eur. J. 21 (2015) 3746–3754. [29] S. Zhang, H. Wang, M. Li, J. Han, X. Liu, J. Gong, Chem. Sci. 8 (2017) 4489–4496. [30] S. Zhang, H. Wang, L. Tang, M. Li, J. Tian, Y. Cui, J. Han, X. Zhu, X. Liu, Appl. Catal. B 220 (2018) 303–313. [31] S. Zhang, X. Liu, M. Li, Y. Wei, G. Zhang, J. Han, X. Zhu, Q. Ge, H. Wang, Catal. Today 324 (2019) 59–65. [32] S. Zhang, H. Wang, M. Li, J. Han, S. Inagaki, X. Liu, Dalton Trans. 46 (2017) 9369–9374. [33] S. Zhang, M. Li, Q. Wu, H. Yang, J. Han, H. Wang, X. Liu, Appl. Catal. B 236 (2018) 466–474. [34] X. Liu, X.B. Li, Z. Guan, J. Liu, J. Zhao, Y. Yang, Q.H. Yang, Chem. Commun. 47 (2011) 8073–8075. [35] O.S. Wenger, Coord. Chem. Rev. 282–283 (2015) 150–158. [36] C.J. Gagliardi, B.C. Westlake, C.A. Kent, J.J. Paul, J.M. Papanikolas, T.J. Meyer, Coord. Chem. Rev. 254 (2010) 2459–2471. [37] M.L. Clark, P.L. Cheung, M. Lessio, E.A. Carter, C.P. Kubiak, ACS Catal. 8 (2018) 2021–2029. [38] J. Zhao, Q. Wang, C. Sun, T. Zheng, L. Yan, M. Li, K. Shao, X. Wang, Z. Su, J. Mater. Chem. A 5 (2017) 12498–12505. [39] Y. Gao, L. Ye, S. Cao, H. Chen, Y. Yao, J. Jiang, L. Sun, ACS Sustain. Chem. Eng. 6 (2017) 781–786. [40] C. Yang, Q. Li, Y. Xia, K. Lv, M. Li, Appl. Surf. Sci. 464 (2019) 388–395. [41] J. Wang, T. Xia, L. Wang, X. Zheng, Z. Qi, C. Gao, J. Zhu, Z. Li, H. Xu, Y. Xiong, Angew. Chem. Int. Ed. 57 (2018) 16447–16689. [42] W. Su, J. Zhang, Z. Feng, T. Chen, P. Ying, C. Li, J. Phys. Chem. C 112 (2008) 7710–7716. [43] J. Wu, X. Li, W. Shi, P. Ling, Y. Sun, X. Jiao, S. Gao, L. Liang, J. Xu, W. Yan, C. Wang, Y. Xie, Angew. Chem. Int. Ed. 57 (2018) 8719–8723. [44] G. Liu, X. Meng, H. Zhang, G. Zhao, H. Pang, T. Wang, P. Li, T. Kako, J. Ye, Angew. Chem. Int. Ed. 56 (2017) 5570–5574. [45] P. Kurz, B. Probst, B. Spingler, R. Alberto, Eur. J. Inorg. Chem. 2006 (2006) 2966–2974.
9
Contents lists available at ScienceDirect
Applied Catalysis B: Environmental journal homepage: www.elsevier.com/locate/apcatb
Heterogeneous molecular rhenium catalyst for CO2 photoreduction with high activity and tailored selectivity in an aqueous solution
T
Shengbo Zhanga,b, Mei Lib, Wenjun Qiub, Jinyu Hanb, Hua Wangb, Xiao Liua,
⁎
a
Key Laboratory of Pesticide & Chemical Biology of Ministry of Education, College of Chemistry, Central China Normal University, Wuhan 430079, China Key Laboratory for Green Chemical Technology, School of Chemical Engineering and Technology, Tianjin University, Collaborative Innovation Center of Chemical Science and Engineering, Tianjin 300072, China
b
ARTICLE INFO
ABSTRACT
Keywords: CO2 photoreduction Heterogeneous molecular catalyst Organosilica nanotubes Microenvironment regulation Aqueous solution
The effective molecular-catalyst-based device that converts CO2 to value-added products is of significance but challenging. Herein, utilizing organosilica nanotubes as the solid chelating ligands, we successfully construct heterogeneous molecular rhenium catalysts for efficient CO2 photoreduction to CO. The nanotube framework compositions and the microenvironment of the rhenium catalysts are finely regulated, aiming to enhance the catalytic selectivity and activity in water-containing systems. By adjusting bipyridine amounts and capping the surface silanols, the preferential adsorption for CO2 over H2O around the active sites is realized. Relative to that of the unmodified catalyst, the CO selectivity increased greatly from 53% to 94% in water/acetonitrile. Furthermore, the heterogeneous molecular catalysts exhibit a total turnover number that is nine times higher than the corresponding homogeneous one (134 versus 15). Using in situ infrared spectroscopy and density functional theory calculations, a reasonable mechanism for high CO selectivity in the presence of H2O is demonstrated.
1. Introduction The photocatalytic reduction of CO2, which mimics natural photosynthesis, aims to achieve efficient light-to-chemical transformation [1–5]. The most serious challenge in this process is the activation of CO2 molecules because of its large potential of the CO2/CO2∙− redox couple (−1.9 V versusthe normal hydrogen electrode (NHE)). An effective alternative to activate CO2 is the proton-assisted multielectron catalytic processes, such as CO2 + 2H+ + 2e− = CO + H2O (E’ = −0.53 V), which requires much less energy [6–11]. Considering the multiple accessible redox states, transition-metal complexes could be utilized as catalysts to minimize the overpotentials for the efficient reduction of CO2. Given this, the device designs for artificial photosynthesis (AP) based on molecular catalysts have always been attracted much research attention [12–15]. However, the molecule-assembled AP systems usually perform low stability and activity. Furthermore, photocatalytic CO2 reduction using molecular catalysts is usually confined to non-aqueous solvent (CH3CN or DMF) systems, which is mainly because that reducing a proton to H2 is more thermodynamically favorable than reducing CO2 in water [16–19]. Given that a water-containing system is necessary to construct AP devices that couples CO2 reduction and water oxidation, designing and developing molecular catalysts with
⁎
high selectivity and stability for the photocatalytic CO2 reduction in a water-containing system are of high significance. The incorporation of metal complex catalysts onto a solid matrix, such as metal organic frameworks (MOFs), covalent organic frameworks (COFs) and periodic mesoporous organosilicas (PMOs) might be a promising way to design AP devices for efficient CO2 photoreduction [20–28]. On the one hand, heterogenization can make active sites spatial isolation and prevent the molecular catalyst deactivation. On the other hand, the matrix framework compositions could be adjusted to preferentially adsorb CO2 rather than water molecules, thus tailoring the catalytic activity and selectivity in water-containing systems. Recently, chelating-ligand-incorporated organosilica nanotubes were prepared from bridged organosilane precursors for the efficient heterogeneous molecular catalyst design in our group [29–34]. These nanotubes with mesoporous diameters have distinct advantages, including various organic functionalities in the nanotube frameworks, high surface areas, fast diffusion of substrates in the tubes, and tunable hydrophilicity/hydrophobicity. Herein, using benzene-bridged organosilica nanotubes containing bipyridine ligands (BPy-NT) as the platform, we report the direct construction of a rhenium complex on the nanotube walls (Re–BPy-NT) through metal complexation for photocatalytic reduction of CO2. By carefully adjusting the bipyridine
Corresponding author. E-mail address: [email protected] (X. Liu).
https://doi.org/10.1016/j.apcatb.2019.118113 Received 14 June 2019; Received in revised form 14 August 2019; Accepted 22 August 2019 Available online 23 August 2019 0926-3373/ © 2019 Elsevier B.V. All rights reserved.
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
4800 scanning electron microscope (SEM, 5 kV) equipped with the Thermo Scientific energy-dispersion X-ray fluorescence analyser. N2 adsorption-desorption isotherms were tested at −196 °C on a Micromeritics Tristar 3000 instrument. The specific surface area was obtained by the Brunauer–Emmett–Teller method; the pore size distribution was obtained by the Barrett–Joyner–Halenda method using the adsorption branch. Carbon, hydrogen and nitrogen contents were determined viaCHN elemental analysis (vario EL CUBE). UV/vis absorption and diffuse reflectance spectra were obtained using Instant Spec BWS003 spectrometers. Re LIII-edge extended X-ray absorption fine structure (EXAFS) spectra were collected at the beamline 1W1B of Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences. The beamline was equipped with a Si (111) double crystal monochromator. The k3-weighted EXAFS oscillations were Fourier transformed into R-space. Transient absorption spectroscopy (TAS) was carried out employing a Femtosecondscale laser flash photolysis spectrometer (Applied Photophysics) equipped with the proper diffuse reflectance accessory. The samples were placed in a closed quartz cuvette, purged with N2 and were excited with a 380 nm laser beam. FT-IR spectra were collected with a Bruker Vertex 80v. In situ IR spectra of a acetonitrile-water-triethanolamine (0.875:1:0.125, v/v) solution containing Re catalyst during photoirradiation (λ > 420 nm) under a CO2 atmosphere. Solid-state 13C cross polarization magic-angle spinning (CP MAS) NMR and 29Si magic angle spinning (MAS) NMR spectra were collected on 400 MHz instruments. Photoluminescence spectra (PL) were done on Jobin Yvon Fluorolog3-21 spectrophotometer at an excitation wavelength of 380 nm by using the powder sample. Photoelectric properties measurements were carried out with an Autolab 302 N electrochemical system with a 300 W xenon lamp and used an Ag/AgCl (0.5 M Na2SO4) as the reference electrode and a Pt plate as the counter electrode. The photocurrent density vs time (I-t) curves of the prepared photoelectrodes were operated under an alternate light irradiation (light on/off cycles: 30 s). Mott–Schottky measurements of the catalysts were conducted in an Autolab 302 N electrochemical system at frequencies of 500, 1000, and 1500 Hz. In Mott–Schottky measures, a three-electrode configuration was used with Pt plate as counter electrode, Ag/AgCl electrode as reference electrode and catalysts on FTO glass substrate as working electrode in 0.5 M Na2SO4 electrolyte. Water contact angles were measured on the contact angle system OCA 20 by pressing the powder sample into a thin sheet.
Scheme 1. Synthetic routes for the design of heterogeneous molecular Rebased catalysts on organosilica nanotubes (Re-BPy0.3-NT-Me).
contents in the framework, the heterogeneous molecular-catalyst-based nanotubes (Re–BPy0.3–NT) could not only facilitate the rapid diffusion of reactants and products but also increase the adsorption capacity of CO2. In order to further inhibit protons from reaching the active sites to produce H2, the remaining accessible surface silanols groups (Si−OH) of the organosilica nanotubes were capped using trimethylsilyl [–Si (CH3)3] (Re–BPy0.3–NT–Me) to increase the hydrophobicity (Scheme 1). The as-prepared Re–BPy0.3–NT–Me shows higher catalytic activity, CO selectivity, and a turnover number (TON) compared with those of the unmodified catalyst and even the corresponding homogeneous one in the CO2 photoreduction system with a high concentration of water. 2. Experimental section 2.1. Synthesis of Re-BPy0.3-NT-Me The rhenium-immobilized nanotube catalysts were prepared by adding BPy-NT (300 mg) to a solution of Re(CO)5Cl (40 mg) in 60 mL of toluene under nitrogen atmosphere. After the suspension was stirred under refluxing conditions for 24 h, the solid phase was obtained by filtration and washed with acetone. The dried samples were named as Re-BPy0.3-NT. Then, 2.0 g of Re-BPy0.3-NT and 8 mL of [(CH3)3Si)]2NH in 30 mL of dry toluene were stirred at room temperature overnight. The resultant powder was filtered, washed and dried under vacuum overnight at 60 °C. The dried samples were named as Re-BPy0.3-NT-Me.
2.4. DFT theoretical calculation All the calculations were used spin-polarized density functional theory (DFT) with generalized gradient approximation (GGA) for exchange-correlation potential embedded in the Vienna Ab Initio Simulation Package (VASP). The ion-electron interaction was described with projector augmented wave (PAW) method. The cutoff energy of the plane-wave basis set was set at 500 eV. The cluster with periodic boundary condition was modeled with a large super cell of 28 × 28 × 28 Å3, eliminating the artificial interaction between molecular electrocatalysts in adjacent cells. The van der Waals (vdW) dispersion by employing the D3 method of Grimme was considered for all the calculation. The Brillouin zone integrations were performed by using Gamma 1 × 1×1 for geometric optimization. The convergence thresholds for structural optimization was set at 0.02 eV/Å in force. The convergence criterion for energy was 10−5 eV.
2.2. Re-BPy0.3-NT-Me-photocatalyzed CO2 reduction reaction In a typical experiment, Re-BPy-NT-Me (2.9 μmol Re) was added to a fresh 40 mL mixed solution containing 17.5 mL of acetonitrile, 20 mL of water and 2.5 mL of triethanolamine. This system was purged with CO2 for 1.0 h. The reaction was operated at the room temperature with a visible light (λ > 420 nm) illumination. The gas molecules generated in the head space were periodically analyzed quantitatively by a Bruker 456 gas chromatograph equipped with a thermal conductivity detector (TCD).
3. Results and discussion
2.3. Characterizations
A rhenium complex was prepared, as shown in Scheme 1. First, organosilica nanotubes containing bipyridine ligands in the framework (BPyx-NT, where x corresponds to the molar ratio of precursors 2 in the initial total precursors) was synthesized as our previous report [29,33]. Subsequently, the rhenium complex [Re(CO)5Cl] was anchored on the
Transmission electron microscopys (TEM) were operated on a JEM2100 F system operating at 200 kV. The morphology of these catalysts were analysed by scanning electron microscopy (SEM) using Hitachi S2
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 1. (a) TEM, (b) SEM and (c) EDS mapping images of Re–BPy0.3–NT–Me. The inset graph in (a) shows the nanotube length distribution in ethanol obtained viadynamics light scattering and magnified high-resolution TEM. (d) Aberration-corrected HAADF-STEM and (e–f) magnified aberration-corrected HAADF-STEM and corresponding Re mapping images of Re–BPy0.3–NT–Me.
nanotube walls (denoted by Re–BPy0.3–NT) viathe metal complexation of Re to the bipyridine unit in BPy0.3–NT (Figs. S1–5). Finally, the remaining accessible silanols (Si–OH) on the organosilica nanotubes were capped using trimethylsilyl [–Si(CH3)3] groups. The obtained sample was denoted by Re–BPy0.3–NT–Me. For comparison, the Re complex was also incorporated onto bipyridine-bridging organosilica nanotubes with fewer bipyridine groups, i.e., BPy0.1–NT (denoted by Re–BPy0.1–NT) (Figs. S7–11) and pure silica nanotubes without phenyl groups (denoted by Re–BPy0.1–SNT) (Figs. S13–17). Fig. 1a and b show the transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images of the short nanotubes Re–BPy0.3–NT–Me. The TEM image clearly reveals that these catalysts comprise nanotubes with ˜60 nm lengths and ˜5-7 nm inner diameters. The SEM image further confirms the nanorod morphology. The energydispersive X-ray spectrometry (EDS) mapping analysis reveals that Re is uniformly distributed on the support (Fig. 1c). In order to more intuitively know the distribution of Re atoms, we performed aberrationcorrected, high-angle, annular, dark-field scanning TEM (HAADFSTEM) measurements (Fig. 1d–f) with sub-angstrom resolutions. Two regions clearly appear in these images; broad, curved, dark lines and small bright spots sized less than ˜0.5 nm. The dark regions correspond to the projections of the nanotube walls along the incident electron beam direction, whereas the bright spots indicate that isolated heavier atoms were present, which were Re atoms in this case. Since the bright spots are not concentrated on the outer/inner surfaces of the nanotube walls but are rather uniform throughout the walls, one can conclude that the Re atoms were uniformly distributed along the nanotube walls with single-site states. The inductively coupled plasma (ICP) measurements show that the Re loadings were 0.288, 0.286, 0.294, and 0.291 mmol g−1 for Re–BPy0.1–SNT, Re–BPy0.1–NT, Re–BPy0.3–NT, and Re–BPy0.3–NT–Me, respectively (Table S1). According to the CHN elemental analysis, the bpyWithout coordination/bpyTotal molar ratios for each catalyst were determined to be 42%, 38%, 73%, and 75%, respectively. To further demonstrate the geometric structure of molecular Re on the skeleton of organosilica nanotubes, various spectroscopies were performed on Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl. In contrast with that of BPy0.3–NT, the Fourier transform infrared (FT-IR) spectrum (Fig. 2a) of Re–BPy0.3–NT–Me exhibits two new absorption peaks at approximately 1921 and 2028 cm−1, which were assigned to the vibration of the Re–CO ligands, in accordance with that of homogeneous Re(bpy)(CO)3Cl. The UV/vis diffuse reflectance spectrum (Fig. 2b) of
Re–BPy0.3–NT–Me shows a new absorption band appeared at approximately 400 nm compared with that of BPy0.3–NT, similar to that of homogeneous Re(bpy)(CO)3Cl, which could be assigned to the metal-toligand (i.e., Re-to-bipyridine-ligand) charge transfer (MLCT) [14,24,26]. The X-ray absorption near-edge structure spectroscopies (XANES) show that the white line resonance peak at approximately 10,538 eV, which is assigned to the allowed 2p–5d transition, intensified with the increase in the d-band vacancies resulting from oxidation (Fig. 2c). Thus, the white line absorption peaks of oxidized Re species are more intense than those of the reduced species, which agrees well with the XPS results. Furthermore, more detailed structural information was obtained viaextended X-ray absorption fine structure spectroscopy (EXAFS). In the Fourier transform of the Re LIII-edge extended EXAFS spectra of Re(bpy)(CO)3Cl and Re–BPy0.3–NT–Me (Fig. 2d), three strong peaks appear at approximately 0.165, 0.220, and 0.264 nm, which can be assigned to Re–C, Re–N, and Re–Cl bond lengths, respectively [14,26]. The XANES spectra and EXAFS Fourier transform at the Re LIIIedge show spectral features corresponding to those of the complex Re (bpy)(CO)3Cl, suggesting the successful formation of the desired Re complex on the pore surface of BPy-NT. To prove the hydrophobicity/hydrophilicity of the as-prepared catalysts, water contact angles were measured (Fig. 3a). Organosilica nanotubes with benzene groups in the skeleton exhibit a significantly higher contact angle and thus higher hydrophobicity than pure silicon nanotubes. More importantly, Re–BPy0.3–NT–Me end-capped using trimethylsilyl [–Si(CH3)3] groups shows the strongest hydrophobicity, with a water contact angle of 96°. These results indicate that hydrophilicity/hydrophobicity can be regulated by finely tuning the matrix framework compositions. In addition, gas sorption using water and benzene was measured at 298 K. As shown in Fig. 3b, c, Re–BPy0.1–NT and Re–BPy0.3–NT exhibit lower water adsorption and higher benzene adsorption at P/P0 = 0.97 than Re–BPy0.1–SNT. Clearly, Re–BPy0.3–NT–Me shows the lowest water adsorption (35.3 cm3/g) and highest benzene adsorption (138.9 cm3/g) at P/P0 = 0.97 because of the improved hydrophobicity. In addition, the CO2 uptake performances of all the materials were investigated at 298 K (Fig. 3d). The adsorption capacities of Re–BPy0.3–NT–Me and Re–BPy0.3–NT were 13.2 and 10.7 mL/g, respectively, which were significantly higher than those of Re–BPy0.1–NT and Re–BPy0.1–SNT. The higher CO2 adsorption capacities of 3
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 2. (a) FT-IR and (b) UV/vis diffuse reflectance spectra of BPy0.3–NT, Re–BPy0.3–NT–Me, and Re(bpy)(CO)3Cl. (c) Re Lш-edge XANES spectra, and (d) Fourier transform EXAFS spectra of Re(bpy)(CO)3Cl, Re–BPy0.3–NT–Me and Re–BPy0.3–NT–Me after reaction.
Fig. 3. (a) Contact angle measurements of Re–BPy0.1–SNT, Re–BPy0.1–NT, Re–BPy0.3–NT and Re–BPy0.3–NT–Me at 298 K. (b) Water and (c) benzene sorption measurements and (d) CO2 uptake performance of Re–BPy0.1–SNT, Re–BPy0.1–NT, Re–BPy0.3–NT, and Re–BPy0.3–NT–Me at 298 K. (e) In situ infrared of CO2 for pure 2,2'-bipyridine at 298 K.
4
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 4. (a) PL spectra, (b) Transient photocurrent responses, (c), (d) Transient optical spectra in the presence of a sacrificial electron donor TEOA, (e) Comparison of kinetics for intramolecular charge transfer at 550 nm and (f) Mott-Schottky plots of Re(bpy)(CO)3Cl and Re–BPy0.3–NT–Me.
Re–BPy0.3–NT–Me and Re–BPy0.3–NT could be primarily attributed to the more bipyridine ligands in their frameworks, which can effectively improve the CO2 uptake (Fig. 3e, Table S1). The optical properties of Re–BPy0.3–NT–Me were investigated by photoluminescence (PL) spectroscopy, as shown in Fig. 4a. Re–BPy0.3–NT–Me exhibits a lower PL intensity than homogeneous Re(bpy) (CO)3Cl when irradiated by a 380-nm laser, with a peak centered at 575 nm, which can be attributed to the efficient photoinduced charge transfer process resulting from the inhibition of intermolecular aggregation by the heterogenization of the catalyst. The transient photocurrent response was also performed to study the photogenerated charge transfer and separation efficiency (Fig. 4b). The photocurrent immediately increased when the light irradiation was turned on, while the photocurrent quickly decreased when the light was switched off, which demonstrates the existence of the photoresponse for Re–BPy0.3–NT–Me. The heterogeneous molecular Re–BPy0.3–NT–Me shows a photocurrent density of approximately 5.5 μA/cm2, which is higher than the complex Re(bpy)(CO)3Cl (3.0 μA/cm2), indicating the efficient photogenerated charge transfer and separation in Re–BPy0.3–NT–Me. To further study light absorption and intramolecular charge transfer (ICT) of Re–BPy0.3–NT–Me, the transient absorption (TA) spectroscopies were performed with a 380 nm excitation (Fig. 4c, d). A stronger and broader positive transient signal in the 550–700 nm region assigned to the absorption of excited ICT appears for Re–BPy0.3–NT–Me compared with the complex Re(bpy)(CO)3Cl [20], which may be mainly due to the fact that the heterogeneous molecular complex has longer lifetime in the excited ICT states than Re(bpy)(CO)3Cl. The TA of Re(bpy)(CO)3Cl shows a positive transient signal in the > 725 nm region, which can be assigned to the formation of catalytically inactive Re-dimers [28], whereas Re–BPy0.3–NT–Me did not have this transient signal due to isolated active sites. Furthermore, the corresponding kinetics comparison of ICT for Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl were also performed. As shown in Fig. 4e, the formation (rising component) of the excited ICT is ultrafast with ˜300 fs for both samples. But interestingly, this excited ICT state of Re–BPy0.3–NT–Me exhibits much longer lifetime (t =161.5 ps) than that of the complex Re(bpy)(CO)3Cl (t = 24.8 ps), suggesting that the heterogenization of Re(bpy)(CO)3Cl
could effectively inhibit charge recombination in Re–BPy0.3–NT–Me. In addition, Mott-Schottky measurements were applied to investigate the electronic band positions and corresponding redox abilities of Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl (Fig. 4f). The flat band positions are −1.68 and −1.50 V (vs Ag/AgCl) for Re–BPy0.3–NT–Me and Re(bpy) (CO)3Cl, respectively, according to extrapolation of C2− values at frequencies of 500, 1000, and 1500 Hz. The positive slopes reveal n-typelike conductivities of Re–BPy0.3–NT–Me and Re(bpy)(CO)3Cl. For ntype semiconductors, the flat band potential is generally approximated by the conduction band potential. Obviously, both samples show more negative reduction potentials (−1.48 V vs NHE for Re–BPy0.3–NT–Me and −1.30 V vs NHE for Re(bpy)(CO)3Cl) than CO2/CO (−0.53 V vs NHE), suggesting the feasibility for photoreduction of CO2 to CO. Subsequently, the performances (Fig. S18) of these Re-based catalysts for the photochemical reduction of CO2 were investigated at room temperature with visible light (λ > 420 nm) illumination in a CO2–saturated acetonitrile–water solution (VMeCN/VH2O = 0.875/1). Fig. 5a shows the CO-TON, H2-TON and CO selectivity with different catalysts under similar reaction conditions at 8 h. No other liquid phase products were detected (Fig. S19). The CO-TON (15.20) of Re–BPy0.3–NT was higher than that of Re–BPy0.1–NT (10.72). The higher catalytic activity of Re–BPy0.3–NT could be attributed to a higher number of bipyridine ligands in the framework (increase CO2 adsorption) and the shorter nanotube length (decrease the diffusion path of reactants and products during the reactions). Notably, the Re–BPy0.3–NT–Me with the capped silanol exhibited the highest CO-TON (19.30) and lowest H2-TON (1.23) with the CO producing rate of 700 μmol/g/h. A CO selectivity as high as 94% could be achieved even in the high water-containing system (50% volume ratio, Table 1). This high selectivity can be explained by the precise modulation of the surface physiochemical properties. When the silanol was capped by methyl groups, the high hydrophobicity of Re–BPy0.3–NT–Me could effectively prevent H2O molecules from reaching the catalytic active centers for H+ reduction. In contrast, CO2 molecules can easily approach the Re sites, thus facilitating the CO2 reduction owing to the enhanced role of CO2 in the competition for adsorption sites. To further prove the hydrophobic/hydrophilic effect, the catalyst with pure silica 5
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 5. (a) The CO-TON, H2-TON and CO selectivity of Re–BPy0.1-SNT, Re(bpy)(CO)3Cl, Re–BPy0.1–NT, Re–BPy0.3–NT and Re– BPy0.3–NT–Me after 8 h at 298 K with 20 mL of water and 2.9 μmol Re. (b) Dependence of the QYs of the Re–BPy0.3–NT–Me on the wavelength of monochromatic light irradiation (435, 500 nm) QY = 2nco/nincident photons × 100%. (c) Recycling performance, (d) The total CO-TON and (e) filtration experiment of Re–BPy0.3–NT–Me for the photochemical reduction of CO2 with 1.0 μmol Re. (f) Effect of water on the CO-TON, H2-TON and CO selectivity for the Re–BPy0.3–NT–Me and Re–BPy0.1–SNT at 298 K with 2.9 μmol Re for the photochemical reduction CO2 (the total volume of solution is 40 mL).
Re(bpy)(CO)3Cl was also used to photocatalyze the same reaction. Unfortunately, Re(bpy)(CO)3Cl became inactive after 4 h of reaction (Fig. S18) and showed a lower CO selectivity of 70%, owing to the decomposition via bimolecular pathways to form metal nanoparticles (Figs. S20, S28) [22,25,27,29,32]. These above results demonstrated the excellent catalytic performance of Re–BPy0.3–NT–Me for the CO2–CO reaction owing to the heterogenization of the molecular catalyst and the finely tuned microenvironment around the active species. In addition, the quantum yields (QYs) strongly depend on the intensity of irradiated light (Fig. 5b). In particular, the QYs of Re–BPy0.3–NT–Me can be as high as 0.146% under 435-nm monochromatic light irradiation. After the reaction, Re–BPy0.3–NT–Me was reused for the next run, and the solid catalyst retained high catalytic activity (Fig. 5c). During the fifth recycle, CO-TON was maintained at approximately 22.85, showing a slight decrease. The total TON reached 134 within 40 h (Fig. 5d). In contrast, the homogeneous catalyst showed a large decrease in CO-TON for the first recycle and was fully deactivated after the second reuse (15 of the total TON, Fig. 5d). In addition, the filtration experiment indicated that the catalytic activity came from Re–BPy0.3–NT–Me, and the solid catalyst was truly heterogeneous (Fig. 5e). The structure analysis of Re-BPy0.3-NT-Me after the reaction revealed the molecular nature of the active site without the formation of Re nanoparticles, as well as the stable nanotube frameworks (Fig. 2c, d, Figs. S20–27). To demonstrate the effect of hydrophobicity/hydrophilicity on the reaction pathways, the effect of the amount of water on the CO-TON,
Table 1 Comparison of catalyst properties during the photochemical reduction CO2.a Entry
Photocatalyst
H2-TONb
CO-TONc
Selectivity (%)
1 2 3 4 5 6 7 8 9 10 11 12
Re-BPy0.3-NT-Me Re-BPy0.3-NT Re-BPy0.1-NT Re-BPy0.1-SNT Re(bpy)(CO)3Cl – BPy0.3-NT Re-BPy0.3-NT-Me/ no CO2 Re-BPy0.3-NT-Me/ no light Re-BPy0.3-NT-Me/no TEOA Re-BPy0.3-NT-Me/2.0 μmol Re Re-BPy0.3-NT-Me/1.0 μmol Re
1.23 2.44 3.16 5.56 4.18 0 0 1.41 0 0 1.48 1.94
19.30 15.20 10.72 6.32 9.64 0 0 0 0 0 24.21 30.60
94 86 77 53 70 – – 0 – – 94 94
a
Reaction conditions: 2.9 μmol of Re, 2.5 mL of TEOA, 17.5 mL of CH3CN, 20 mL of H2O, and λ > 420 nm, at room temperature. b H2-TON is defined as the number of evolved hydrogen molecules per catalytic site (Re). c CO-TON is defined as the number of evolved CO molecules per catalytic site (Re).
nanotubes but no phenyl groups (Re–BPy0.1–SNT) was used to photocatalyze the CO2–CO reaction under the same reaction conditions. It can be seen that Re–BPy0.1–SNT exhibited the lowest CO-TON (6.32) and highest H2-TON (5.56), with a CO selectivity of 53% (Fig. 5a), which could be mainly attributed to the high hydrophilicity of the pure silica nanotube framework. For comparison, the homogeneous catalyst 6
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 6. (a) CVs of Re–BPy0.3–NT–Me in a 0.1-M NBu4PF6 CH3CN/H2O (v/v = 0.875:1) solution with 1 mM Re catalysts under a saturated N2 (black) and CO2 (red) atmosphere using a 3mm glassy carbon electrode and Ag/AgCl electrode at a scan rate of 100 mV s−1. (b) In situ FT-IR spectra of CO2 photoreduction over Re–BPy0.3–NT–Me recorded after various irradiation times.
the ligand exchange of the molecular CO of species E with Cl− produces the initial Re complex A to complete the CO2 reduction pathway (1). Evidently, adding water plays a proton-promoting role. Simultaneously, protons are reduced to form H2 (pathway (2)) during this process, which is the main reaction competing with the CO2 reduction [17–19,45]. Interestingly, for hydrophobic Re–BPy0.3–NT–Me, more CO2 and fewer protons are allowed to reach the active site to react; thus, water primarily play a proton-promoting role. However, for hydrophilic Re–BPy0.1–SNT, excess protons can reach the active sites, resulting in more competition between the two reactions (1 and 2), thus weakening the reduction of CO2 (1). In addition, density functional theory (DFT) calculations were employed to investigate the binding energies of H2O onto the Re–BPy0.3–NT and Re–BPy0.3–NT–Me. As shown in Fig. 7b, the binding energy of H2O on Re–BPy0.3–NT is −0.48 eV, which is much stronger than that of Re–BPy0.3–NT–Me (−0.28 eV). Further, such an obvious difference in the binding energy of H2O over Re–BPy0.3–NT and Re–BPy0.3–NT–Me has been proved by experimentally (Fig. 3a, b, and c). Notably, the weaker binding energy of H2O on Re–BPy0.3–NT–Me can suppress the undesired metal–H2O adduct F with H2 as a by-product. Finally, the evolution of the free energy of CO2 photoreduction to CO on the Re–BPy0.3–NT–Me was calculated using DFT to better evaluate and understand the reaction processes and related active species (Fig. 7c). As shown in Fig. 7c, the pathway for the reduction of CO2 to CO viaa Re–carboxyl (Re–COOH*) intermediate viathe transfer of two protons and electrons was considered. Initially, the CO2 molecule is adsorbed on the Re sites, with concerted protonation and electron transfer, leading to the formation of the Re–COOH* intermediate. Subsequently, this intermediate can be further reduced viaproton–electron transfer to form a Re–CO* adduct, releasing one H2O molecule. Here, it should be mentioned that the hydrophobic material Re–BPy0.3–NT–Me facilitates the separation of water from the reaction system, thereby accelerating the reaction. The main product CO is finally generated from the ligand exchange of the molecular CO of species E with Cl−. We find that the formation of the Re–COOH* species viaprotonation (energy barrier of 0.39 eV) and the release of molecular CO viathe ligand exchange (energy barrier of 0.46 eV) are the two main rate-limiting steps. Importantly, adding weak Bronsted acids (H2O) can effectively facilitate these catalytic pathways. On the one hand, adding water provides more protons to accelerate the formation of the Re–COOH* intermediate viathe protonation of CO2. On the other hand, the protons H+ from the H2O can further promote the conversion of ReOH* to Re–CO*. As a result, the concentration of Re–CO* will continue to increase, thereby accelerating the release of CO according to reaction equilibrium theory. However, the DFT calculations of the free energy evolution of protons reducing to H2 indicate that reducing protons is easier than reducing CO2 in this system (Fig. S31), which is the main competitive reaction for CO2 reduction. In response to this problem, adjusting the microenvironment around the active sites will provide
H2-TON and CO selectivity in this reaction system was investigated using Re–BPy0.3–NT–Me and Re–BPy0.1–SNT as control catalysts (Fig. 5f). Surprisingly, adding water can improve the catalytic performance of the hydrophobic Re–BPy0.3–NT–Me catalyst to some extent, mainly due to that the added water could promote the proton-coupled electron transfer [35–40]. Furthermore, the presence of the protoncoupled electron transfer effect was also confirmed by the photocurrent response test (Fig. S29). A high CO selectivity over 94% can be obtained with a high CO-TON over 18 and an almost constant H2-TON below 1.5 in 0–20 mL of water. However, for the hydrophilic Re–BPy0.1–SNT, the CO selectivity and CO-TON always decrease, whereas H2-TON increases with the addition of water. These results indicate that the hydrophobic Re–BPy0.3–NT–Me catalyst can achieve high activity and selectivity in a system with a high concentration of water by fine-tuning the microenvironment around the active species, which provides a promising platform to design devices for artificial photosynthesis. Next, the redox properties of Re–BPy0.3–NT and Re–BPy0.3–NT–Me were investigated by recording cyclic voltammograms (CVs) to further illustrate the effect of hydrophobicity on CO2 reduction (Fig. 6a and Fig. S30). The CV of Re–BPy0.3–NT–Me exhibits increased current at Eonset = −0.74 V under a CO2 atmosphere, and this current density (−1.22 mA/cm2) is more than 2.3 times higher than that under an N2 atmosphere. Interestingly, Re–BPy0.3–NT–Me exhibits a lower Eonset and higher current density than Re–BPy0.3–NT (−0.80 V and −0.94 mA/ cm2, respectively) and Re(bpy)(CO)3Cl (−0.82 V and −0.82 mA/cm2, respectively). These results indicate that hydrophobic Re–BPy0.3–NT–Me is more favorable for CO2 reduction. To further investigate the photocatalytic CO2 reduction process, in situ infrared spectroscopy was conducted for the Re–BPy0.3–NT–Me sample (Fig. 6b). Some new peaks clearly emerge with the increased irradiation time from 0 to 64 min; the peaks at 1297, 1473, and 1650 cm−1 are assigned to the Re–COOH groups [41–44]. The intensity of these peaks increased with increasing exposure time, suggesting that the Re–COOH groups are intermediates during the photoreduction of CO2 to CO. Subsequently, the Re–COOH species accepts another proton and undergoes dehydration to form a Re–CO complex, which gives rise to some additional peaks at 1870–2050 cm−1. Finally, the CO molecule in Re–CO can be readily released and desorbed from the catalyst surface, as indicated by the peak at 2160 cm−1 [41–44]. Based on the experimental results, a possible mechanism underlying the photocatalytic reduction of CO2 in a system containing such a high concentration of water is proposed in Fig. 7a. First, the initial Re complex A is converted into the one-electron reduced species B after irradiation with visible light based on the MLCT. Then, species B is reductively quenched using TEOA to form species C. Subsequently, the CO2 adduct D is formed by the nucleophilic attack of the Re complex C on the electrophilic carbon atom of CO2, accompanied by coordination with H+ of H2O. The CO2 adduct D can accept another H+ from H2O to form intermediate species E, thus driving the CO2 reduction. Finally, 7
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al.
Fig. 7. (a) Possible photocatalytic CO2 reduction mechanism. DFT calculations: (b) the binding energy of H2O onto Re–BPy0.3–NT and Re–BPy0.3–NT–Me and (c) free energy evolution of the reduction of CO2 to CO on Re–BPy0.3–NT–Me.
obvious advantages. Furthermore, it must be mentioned that the desorption of CO intermediates on the molecular catalyst is also very important for photoreduction CO2 to CO [13,18,20]. Through the hydrophobicity regulation, we hope to reduce the competitive reaction of water, and also hope not to reduce the desorption process of the product CO. Fortunately, the DFT results show that binding energy of ReCO on the Re-BPy0.3-NT-Me and Re-BPy0.3-NT is −1.15 eV and −1.08 eV (Fig. S32), respectively, which indicates that the desorption of product CO on the molecular catalyst is relatively similar. Therefore, we believe that the improvement of CO2 reduction performance for the modified Re-BPy0.3-NT-Me material is mainly by reducing the competitive reaction of H2O according to the DFT and experimental results. For the optimized Re–BPy0.3–NT–Me, such a microenvironment can not only prevent excessive protons from reacting on the active site owing to high hydrophobicity but also effectively increase the adsorption capacity of CO2 owing to more bipyridine on the skeleton, thus playing a dual role.
intermediate was Re–COOH* viathe protonation of CO2 after modifying the microenvironment around the active sites. We believe that this work will provide a new guideline for the device design using functional organosilica nanotubes as highly tunable one-dimensional catalytic materials for artificial photosynthesis. Declaration of Competing Interest The authors declare no competing financial interests. Acknowledgments We acknowledge the National Natural Science Foundation of China (U1662109) for financial support. The Program of Introducing Talents of Discipline to Universities of China (111 program, < GN2 > B17019) < /GN2 > is also acknowledged. Appendix B. Supplementary data
4. Conclusion
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.apcatb.2019.118113.
In conclusion, we have constructed efficient heterogeneous molecular rhenium catalysts based on organosilica nanotubes and proposed a microenvironment modification strategy to tune the surface compositions for the highly selective photoreduction of CO2 in aqueous solutions. By adjusting the bipyridine contents in the framework of organosilica nanotubes and subsequently capping the surface silanols, we modulated the CO2 adsorption capacity and hydrophobicity/hydrophilicity around the active sites, leading to a high CO selectivity of 94% in a photocatalytic CO2 reduction system with a high concentration of water. DFT calculations and in situ infrared spectroscopy reveal that the binding energy of H2O decreased considerably, and the main
References [1] X. Liu, S. Inagaki, J. Gong, Angew. Chem. Int. Ed. 55 (2016) 14924–14950. [2] N. Elgrishi, M.B. Chambers, X. Wang, M. Fontecave, Chem. Soc. Rev. 46 (2017) 761–796. [3] S. Berardi, S. Drouet, L. Francas, C. Gimbert-Surinach, M. Guttentag, C. Richmond, T. Stoll, A. Llobet, Chem. Soc. Rev. 43 (2014) 7501–7519. [4] X. Chang, T. Wang, J. Gong, Energy Environ. Sci. 9 (2016) 2177–2196. [5] X. Liu, F. Kang, C. Hu, L. Wang, Z. Xu, D. Zheng, W. Gong, Y. Lu, Y. Ma, J. Wang, Nat. Chem. 10 (2018) 1201–1206. [6] M. Aresta, A. Dibenedetto, A. Angelini, Chem. Rev. 114 (2014) 1709–1742.
8
Applied Catalysis B: Environmental 259 (2019) 118113
S. Zhang, et al. [7] E.E. Benson, C.P. Kubiak, A.J. Sathrum, J.M. Smieja, Chem. Soc. Rev. 38 (2009) 89–99. [8] K.A. Grice, Coord. Chem. Rev. 336 (2017) 78–95. [9] A.J. Morris, G.J. Meyer, E. Fujita, Acc. Chem. Res. 42 (2009) 1983–1994. [10] M.L. Clark, A. Ge, P.E. Videla, B. Rudshteyn, C.J. Miller, J. Song, V.S. Batista, T. Lian, C.P. Kubiak, J. Am. Chem. Soc. 140 (2018) 17643–17655. [11] B. Reuillard, K.H. Ly, T.E. Rosser, M.F. Kuehnel, I. Zebger, E. Reisner, J. Am. Chem. Soc. 139 (2017) 14425–14435. [12] K.M. Choi, D. Kim, B. Rungtaweevoranit, C.A. Trickett, J.T. Barmanbek, A.S. Alshammari, P. Yang, O.M. Yaghi, J. Am. Chem. Soc. 139 (2017) 356–362. [13] B. Zhang, L. Sun, Chem. Soc. Rev. 48 (2019) 2216–2264. [14] M. Waki, K.I. Yamanaka, S. Shirai, Y. Maegawa, Y. Goto, Y. Yamada, S. Inagaki, Chem. Eur. J. 24 (2018) 3846–3853. [15] R. Kuriki, M. Yamamoto, K. Higuchi, Y. Yamamoto, M. Akatsuka, D. Lu, S. Yagi, T. Yoshida, O. Ishitani, K. Maeda, Angew. Chem. Int. Ed. 56 (2017) 4867–4871. [16] S. Fukuzumi, Y.M. Lee, H.S. Ahn, W. Nam, Chem. Sci. 9 (2018) 6017–6034. [17] S. Sato, T. Morikawa, T. Kajino, O. Ishitani, Angew. Chem. Int. Ed. 52 (2013) 988–992. [18] T. Ouyang, H.-H. Huang, J.-W. Wang, D.-C. Zhong, T.-B. Lu, Angew. Chem. Int. Ed. 56 (2017) 738–743. [19] M. Liu, Y.-F. Mu, S. Yao, S. Guo, X.-W. Guo, Z.-M. Zhang, T.-B. Lu, Appl. Catal. B 245 (2019) 496–501. [20] S. Yang, W. Hu, X. Zhang, P. He, B. Pattengale, C. Liu, M. Cendejas, I. Hermans, X. Zhang, J. Zhang, J. Huang, J. Am. Chem. Soc. 140 (2018) 14614–14618. [21] Y. Wang, N.Y. Huang, J.Q. Shen, P.Q. Liao, X.M. Chen, J.P. Zhang, J. Am. Chem. Soc. 140 (2018) 38–41. [22] X. Liu, Y. Maegawa, Y. Goto, K. Hara, S. Inagaki, Angew. Chem. Int. Ed 55 (2016) 7943–7947. [23] G. Jiang, K. Geng, Y. Wu, Y. Han, X. Shen, Appl. Catal. B 227 (2018) 366–375. [24] H.-P. Liang, A. Acharjya, D.A. Anito, S. Vogl, T.-X. Wang, A. Thomas, B.-H. Han, ACS Catal. 9 (2019) 3959–3968. [25] Y. Ueda, H. Takeda, T. Yui, K. Koike, Y. Goto, S. Inagaki, O. Ishitani, ChemSusChem 8 (2015) 439–442. [26] M. Wen, K. Mori, Y. Kuwahara, H. Yamashita, ChemCatChem 7 (2015) 3519–3525.
[27] C. Wang, Z. Xie, K.E. deKrafft, W. Lin, J. Am. Chem. Soc. 133 (2011) 13445–13454. [28] C.D. Windle, E. Pastor, A. Reynal, A.C. Whitwood, Y. Vaynzof, J.R. Durrant, R.N. Perutz, E. Reisner, Chem. Eur. J. 21 (2015) 3746–3754. [29] S. Zhang, H. Wang, M. Li, J. Han, X. Liu, J. Gong, Chem. Sci. 8 (2017) 4489–4496. [30] S. Zhang, H. Wang, L. Tang, M. Li, J. Tian, Y. Cui, J. Han, X. Zhu, X. Liu, Appl. Catal. B 220 (2018) 303–313. [31] S. Zhang, X. Liu, M. Li, Y. Wei, G. Zhang, J. Han, X. Zhu, Q. Ge, H. Wang, Catal. Today 324 (2019) 59–65. [32] S. Zhang, H. Wang, M. Li, J. Han, S. Inagaki, X. Liu, Dalton Trans. 46 (2017) 9369–9374. [33] S. Zhang, M. Li, Q. Wu, H. Yang, J. Han, H. Wang, X. Liu, Appl. Catal. B 236 (2018) 466–474. [34] X. Liu, X.B. Li, Z. Guan, J. Liu, J. Zhao, Y. Yang, Q.H. Yang, Chem. Commun. 47 (2011) 8073–8075. [35] O.S. Wenger, Coord. Chem. Rev. 282–283 (2015) 150–158. [36] C.J. Gagliardi, B.C. Westlake, C.A. Kent, J.J. Paul, J.M. Papanikolas, T.J. Meyer, Coord. Chem. Rev. 254 (2010) 2459–2471. [37] M.L. Clark, P.L. Cheung, M. Lessio, E.A. Carter, C.P. Kubiak, ACS Catal. 8 (2018) 2021–2029. [38] J. Zhao, Q. Wang, C. Sun, T. Zheng, L. Yan, M. Li, K. Shao, X. Wang, Z. Su, J. Mater. Chem. A 5 (2017) 12498–12505. [39] Y. Gao, L. Ye, S. Cao, H. Chen, Y. Yao, J. Jiang, L. Sun, ACS Sustain. Chem. Eng. 6 (2017) 781–786. [40] C. Yang, Q. Li, Y. Xia, K. Lv, M. Li, Appl. Surf. Sci. 464 (2019) 388–395. [41] J. Wang, T. Xia, L. Wang, X. Zheng, Z. Qi, C. Gao, J. Zhu, Z. Li, H. Xu, Y. Xiong, Angew. Chem. Int. Ed. 57 (2018) 16447–16689. [42] W. Su, J. Zhang, Z. Feng, T. Chen, P. Ying, C. Li, J. Phys. Chem. C 112 (2008) 7710–7716. [43] J. Wu, X. Li, W. Shi, P. Ling, Y. Sun, X. Jiao, S. Gao, L. Liang, J. Xu, W. Yan, C. Wang, Y. Xie, Angew. Chem. Int. Ed. 57 (2018) 8719–8723. [44] G. Liu, X. Meng, H. Zhang, G. Zhao, H. Pang, T. Wang, P. Li, T. Kako, J. Ye, Angew. Chem. Int. Ed. 56 (2017) 5570–5574. [45] P. Kurz, B. Probst, B. Spingler, R. Alberto, Eur. J. Inorg. Chem. 2006 (2006) 2966–2974.
9